On Aug 13, 2012, at 09:01 , Kyle Sluder <k...@ksluder.com> wrote: > As was discussed on the objc-language list recently, the ARC spec > specifically guarantees that an object's lifetime includes the entire > duration of a method invocation. An object *cannot* be deallocated while one > of its methods is executing.
That was only under the assumption that the receiver's pointer was loaded from a weak variable. It'll also be true if the receiver's pointer is loaded from a strong stack variable (because the stack reference's lifetime cannot end before the receiver's method is called -- even under optimization -- and there isn't AFAICT any subsequent way for it to end until the receiver's method returns). However, if the receiver's pointer is loaded from a *strong* instance variable, there's no memory management applied to it. It's not retained when loaded (which I just verified by looking at a disassembly), and it's not retained on entry to the method itself (the Clang ARC spec explains why not). Setting the strong ivar to nil during the method execution would be a really bad idea. At least, that's how I understand the situation. _______________________________________________ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
